1. Novel approach for calibration breakdown voltage of large area SiPM
Sergei Dolinsky
Imaging Technologies
GE Global Research Center
Niskayuna, NY
International Workshop on New Photon-Detectors: PhotoDet 2012
June 13-15, 2012 – LAL Orsay, France

2. Geiger Mode APD and Gain
Operation principle: Photodiode-> APD ->Break down -> Geiger I
Break down
G=1
APD
Geiger
VOV V
VBD
VOp
GainGeiger = VOV*Cdiode

3. DC vs. pulsed Gain Calibration
Gain definition
Diode, Gain1
APD, Gain <1000
Geiger Mode APD, Gain ~
Gain 1,
QE measurement
Linear mode:
QE and Gain(V)
Geiger mode, with limited Ncells
PDE(Vov), Gain(Vov) DC OK
OK,
dark current contribution is small
Difficult:
After pulses and Crosstalk
Pulsed mode
Slow pulse, low noise, low bandwidth amplifier
Medium intensity light pulse
~1 SPE per pulse. Need high gain, low noise amplifier
Linear - PMT, PIN diode or APD
Signal = NPh*QE*Gain
SiPM
Signal = NPh*QE*PGeiger*GSPE
PGeiger(Vov , Nph/Ncells)
Need high gain, low noise amplifier or different method

4. SPE based calibration Works very well for small devices ~1-10 mm2
Can use non-calibrated light source
Need for PDE vs. Vov measurements
Need high gain, low noise amplifier
MPPC – S C SN#407, 3x3 mm2
MPPC – S C SN#407, 3x3 mm2
Laser 405 nm PicoQuant LDH-D-C-405, MHz
10 kHZ , attenuated OD 4.0
Preamp. 50 Ohm,
Oscilloscope LeCroy WaveRuner 610Zi 1 GHz , filter -BW 20 MHZ

5. SPE based calibration
Works very well for small devices ~1-10 mm2 with low noise
Need high gain, low noise amplifier
Problematic for big devices ~ mm2 : Long recharging time, high dark count rate, crosstalk and after pulses make it difficult
MPPC – S C SN#407, 4 x 3x3 mm2
3x3 mm2 -> 6x6 mm2
MPPC-50um intrinsic impedance
35 Ohm+320 pF -> Ohm+1500 pF
Recovery time
 = CSiPM*(RSiPM+RInput)
 = 27 ns >  = 75 Ohm
1 SPE, Vov=1 V @50 Ohm
Vpeak = 165 uV -> Vpeak = 60 uV.

6. Average pulse based calibration
Measure average pulse amplitude and charge vs. Vbias for ~ SPE
Fast method
Pulse shape is the same as SPE pulse shape
Measure PDE*GSPE *(1+Xtalk)
Non-linear behavior in Vbr region -> difficult to interpolate
Afterpulses
Crosstalk

7. “High intensity” pulsed calibration
Let’s trigger all cells by HUGE light pulse
PDE(Nph)~ 1-exp(-PGeiger*Nph/Ncells)
No PDE dependence
Signal is linear for small Vov

8. “High intensity” pulsed calibration
Optical power requirements:
Epulse = Eph(λ)*Ne-h/QE*Ncells*AreaIlluminated/AreaSiPM
For 1 cm2 illuminated area and 50 um u-cell
Epulse(405 nm) =1.2 pJ
1 ns pulse PeakPowerOptical (405) = 1 mW
PicoQuant LDH-D-405 ~50-60 pJ /pulse
Laser Diodes ~100mW available
Use SiPM below Vbr to measure pulse intensity
I(G=1) = Frep*Ncells*Ne-h*Qe =10 kHz*3600*30*1.6e-19 = 180 pA

9. Dark current measurement
Keithley 2400 SourceMeter I-V measurements:
How fast to do I-V scan?
Remove C from Vbias ( RC, leakage ~nA) !
SiPM stabilization time ~10 sec
VBias
Rload
SiPM
Ampl.
VLED A
Keithley 2400

10. “High intensity” pulsed calibration
MPPC – S C SN#407, 3x3 mm2
Laser 405 nm PicoQuant LDH-D-C-405, MHz
10 kHZ , attenuated OD
For wide range of light intensity I-V has linear part

11. “High intensity” pulsed calibration
Gain/dV = dI/dV/Frep/Ncell = Cdiode
I diode
e-h/ucell
Gain/dV
Vbr
No filter
4.5 nA
800
152 fF
69.76
ND 0.3
2.4 nA
420
145 fF
ND 0.7
0.8 nA
140
132 fF
69.78
ND 1.0
0.28 nA 50
122 fF
69.85
ND 1.5
0.07 nA 12
97 fF
70.05

12. “High intensity” pulsed calibration
50 Ohm = Vov* 50/(RSiPM +50)
e-h/ucell
Slope
Gain/dV
No filter
800
0.9
152 fF
ND 0.7
140
0.75
132 fF
ND 1.5 12
0.6
97 fF

13. “High intensity” pulsed calibration HPK 6x6 mm2
50 Ohm = Vov* 50/(RSiPM +50)
e-h/ucell
Slope
Gain/dV
Laser spot ~5 cm2 30
1.0
129 fF

14. “High intensity” pulsed calibration SensL 6x6 mm2
50 Ohm = Vov* 50/(RSiPM +50)
SensL Micro SM X13 Lot
e-h/ucell
Slope
Gain/dV
Laser spot ~5 cm2 10
0.58
158 fF

15. Gain = (I – Idark)/(I27V-Idark,27V)
SiPM as APD
Gain = (I – Idark)/(I27V-Idark,27V)
SiPM works as APD
Limited linearity
Gain ~1000 at Vbr

16. Conclusion
The High Intensity pulse calibration method is simple and reproduces Vbr measured by SPE
I-V measurement can be used during SiPM production
No need of a high gain amplifier, can be used with FE electronics optimized for scintillator pulses
SiPM can operate as an APD below break down
Simple model of SiPM and Geiger avalanche development does not work for very high number of primary e-h